The investigator proposes to bring theoretical modeling from the reaction time domain to bear on the domain of aging and reaction time. Random walk and diffusion models are able to fit a range of experimental data beyond mean reaction time, and there appears to be a consensus that these models are setting new standards in accounting for reaction time phenomena. Five lines of experiments are proposed, each requiring two-choice decisions about simple cognitive and perceptual stimuli. Each kind of experiment places different limits on cognitive processes: the first two require subjects to learn stimulus-to-response mappings; the second two limit perceptual processes by limiting encoding time and stimulus discriminability; and the third requires memory. For all five tasks, performance will be examined with standard reaction time tasks, speed versus accuracy instructions, and deadlines that limit the time subjects have to respond. Models will be fit to all of the data from each task, including reaction times for correct and error responses, response accuracy, the relations among reaction time and accuracy, and the shapes of the reaction time distributions. Parameters of the models will be used to interpret aging deficits. For example, the parameters of the diffusion model allow the quality of the information entering decision processes to be separated from other components of the processes. The investigators anticipate that this theoretical approach will provide a more complete picture of processing than the Brinley plot analyses that currently dominate explanations of aging effects on reaction time.